Adaptive Incoherent Digital Holographic Fast Three-dimensional Imaging

The evolution of microscopy has been driven by life scientists who wish to oberve structure and measure phenomena that were smaller,fainter,faster and deeper inside the tissue than ever before.The researches about three-dimensional(3D)microscopy have been so far focused on the improving of performances such as spatial resolution,imaging field-of-view and imaging speed.Over the past centuries various 3D imaging techniques with high resolution,faster speed and larger field-of-view have been proposed.Among them,incoherent holographic techniques provide the ability of non-scanning 3D imaging for spatial incoherent samples.The high resolution 3D fluorescence imaging achieved by incoherent digital holography has great potential toward in-vivo imaging of biological samples.The spatial structures of biological tissues or cells are usually complex and three-dimensionally.It is important that the imaging skill has sufficient axial imaging resolution and offers the ability to collect serial optical sections from thick specimens.On the other hand,imaging performances of 3D microscope are strongly affected by optical aberrations caused by inhomogeneities of the refractive index in the sample.Developing of a fast,3D,tomographic and adaptive microscopic imaging technique can benefit the the in-vivo imaging of biological samples.However,the insufficient axial resolution and imaging speed of incoherent digital holography limits its further applications in the imaging of living samples.In this thesis,by combing with compressive sensing and parallel phase shifting techniques,the performances such as axial imaging resolution and imaging speed of incoherent digital holography are improved.To improve the 3D imaging performance of incoherent holography in practical application of biological samples,a 3D adaptive imaging technique is also proposed based on fluorescence digital holography.The research described in this thesis can be summarized as:1.The basic principle and imaging performance of incoherent digital holography are introduced.The theoretical models of recording and reconstruction procedure of two implementations of incoherent holography,Triangular holography and Fresnel Incoherent Correlation Holography(FINCH),are investigated in detail.Reconstruction algorithms for different types of holograms are introduced.Theoretical imaging performances of incoherent digital holography are illustrated in FINCH.A brief introduction about the basic principle of compressive sensing is given.The concept of reconstruction guarantees and accuracy is introduced.Basic principles about sensorless adaptive optics,mathematic description of optical aberration and image quality metric are also discussed.2.Single-exposure 3D imaging of samples is achieved in off-axis incoherent triangular holography.3D spatial location of the sample can be extracted from the hologram with a resolution of 0.3mm laterally and 0.35 mm axially.The application of the proposed method is demonstrated by retrieving the 3D moving trajectory of a living zebra fish larva.3.Compressive parallel phase shifting incoherent digital holography is proposed based on on-axis common-path FINCH technique.Comparing with off-axis holographic single-exposure imaging technique,the spatial bandwidth product can be utilized more efficiently in the propsed method.Spatial multiplexing parallel phase shifting technique is implemented during the recording of hologram.Sub-holograms are extracted from single-exposure spatially multiplexed hologram.Compex-valused hologram is then generated using the sub-holograms and single-exposure 3D imaging is achieved.Image-based phase shifting error correction algorithm is proposed to correct the potential phase shifting error in the system.Theoretical models of parallel incoherent digital holography are established in the framework of compressive sensing.Reconstruction error and noise casued by the undersampling of holograms is suppressed using compressive reconstruction.Therefore,the imaging performance of the parallel incoherent digital holography is improved.4.Compressive sensing is combined with FINCH and Interference Coded Aperture Correlation Holography(I-COACH)respectively for non-scanning 3D tomographic imaging.Compressive sensing model of FINCH 3D imaging is established.Theoretical reconstruction accuracy of compressive FINCH is evaluated.Axial reconstruction accuracy of the conventional FINCH is improved by compressive reconstructing the hologram.Non-scanning 3D tomographic imaging ability of the proposed method is demonstrated in experiments.The improvements of compressive FINCH are demonstrated experimentally.Compressive I-COACH is proposed for single-exposure 3D tomographic imaging.Reconstruction accuracy of the conventional COACH is improved by compressive reconstructing the holograms.Background noise and reconstruction error are suppressed effectively.Single-exposure 3D imaging is achieved by the proposed compressive COACH without any sacrifice on imaging field-of-view.Furthermore,the axial imaging performance is also improved.5.Basic concept of incoherent holographic adaptive imaging is demonstrated for optical aberrations correction.The validity of wavefront sensing and correction in guide-star-based incoherent holographic adaptive optics is demonstrated in triangular holography.Effects of the size and axial location of guide star on the aberration correction are discussed.A guide-star-free,sensorless and modulator-free computational adaptive fluorescence digital holography is proposed toward 3D imaging of tissue specimen.In our approach,only one complex-value hologram is sufficient to measure and then correct the aberrations,which resulting in fast acquisition speed,lower exposure time,and the ability to image in three-dimensions without the need to scan the sample or any other element in the system.Inverted fluorescence digital holographic microscope is built.The imaging performances of the proposed method are demonstrated experimentally.We show proof-of-principle experiments on a tissue phantom containing fluorescence particles and demonstrate the 3D imaging and anisotropic aberration correction ability of the proposed method.Furthermore,we present three-dimensional reconstructions of actin-labeled MCF7 breast cancer cells,showing improved resolution after correction of aberrations.Experiments demonstrate the validity of our method and show the great potential of non-scanning adaptive three-dimensional microscopy in imaging biological samples with improved resolution and signal-to-noise ratio.